Dear Alexy,
Inelastic nuclear interactions of alpha particles are modelled in FLUKA. At
your energy, the event generator is already linked (for ion interactions above
125 MeV/n one would have to explicitly link the event generator with ldpmqmd).
In the output file you will see that the inelastic scattering length of
alpha particles in your beryllium at 5 MeV is 11.15 cm. Thus, nuclear
inelastic interactions will be a fairly uncommon event (you would have
to sample many primary particles to eventually attain decent statistics).
Since this is the type of event you are interested in, and since your
geometry is effectively a thin target (electronic stopping will bring
your 5-MeV alphas to a halt at pretty shallow depths) you may want to
consider a LAM-BIAS card to shorten the inelastic scattering length of
your alphas in beryllium. This should allow you to populate your
scorings more efficiently.
Cheers,
Cesc